Ten years ago, a user acquisition agency could charge 15% of media spend and defend it without blinking. On a $500,000 monthly budget that was $75,000 a month, and it bought something real: people building audiences by hand, pulling reports by hand, and making the hundred small targeting decisions that actually moved performance.
That work is largely gone. The fee, in a lot of places, is not.
We charge under 5%, improving as you scale, and we think most agencies still charging 15% are charging for labour the platforms have automated away.
Page Contents
What 15% used to pay for
It is worth being fair to the model, because it was not a scam. It was priced against real work.
Manual targeting. Someone built the audiences: interest stacks, lookalike tiers, exclusion lists, placement by placement. Getting this right was a genuine skill, and the difference between a good and a bad media buyer showed up directly in CPA.
Manual analysis. Someone pulled the numbers, built the pivot, found the segment that was quietly bleeding money, and caught it before the month closed.
Manual production coordination. Every creative was a small project: brief, shoot, edit, revise, traffic. Volume scaled with headcount, because there was no other way for it to scale.
Fifteen percent of a large budget funded a team doing all three. On the biggest accounts it funded quite a large one.
What actually changed
Two things, and they hit from opposite ends.
Algorithmic targeting ate the buying. You do not build the audience any more. You give the platform a budget, a conversion event and a creative, and it finds the people. The manual targeting craft that justified a large share of the fee has been absorbed into the platform, and the platforms are better at it than the humans were. The analysis went the same way: much of what a buyer used to find by hand, the system now acts on before anyone opens a dashboard.
AI production ate the making. Generating a creative no longer costs what it did. Concept to shippable asset has collapsed from a project into an afternoon. The production coordination that used to scale with headcount does not scale with headcount any more.
So the two largest line items inside the old 15% both got dramatically cheaper to deliver. A fee designed around them has not been repriced to match.
What is actually left to pay for
Not nothing. The work moved rather than disappeared, and where it moved is worth paying for:
- Deciding what to make. The strategic bet, the angle, the thing the ad is actually claiming. Automation made execution cheap and made this the bottleneck.
- Taste and curation. When you can generate forty versions of a shot for almost nothing, the scarce resource is the judgment to know which two are worth running.
- Testing discipline. Knowing what to test, how much budget a test needs to be readable, and when to cut. Getting this wrong wastes far more money than a fee ever will. We have written up both halves of it: how many creatives to test per week and when you can actually judge a creative.
- Knowing when the platform is wrong. Algorithmic delivery is very good and occasionally confidently incorrect. Recognising which is which is still a person’s job.
Notice what those have in common: none of them scale linearly with your media budget. Deciding the creative bet for a $300,000 account is not ten times the work of deciding it for a $30,000 one. It is somewhat more work. Not ten times.
That is the whole argument against a flat percentage. The fee scales linearly with spend. The work does not.
What good pricing looks like now
The percentage should fall as spend rises. A rate card charging the same percentage at $30,000 and $300,000 is charging you for a workload that is not there. Ours is under 5% and improves as you scale, comfortably under 5% above $100,000 a month. It also reflects channel count, because each new channel is genuinely more work in a way that more spend on one channel is not.
Or price the work, not the spend. A retainer against a defined scope, so the fee tracks what is produced rather than what is spent. That is where we start, at $7,500 a month for creative plus growth marketing. Full detail: how much a mobile app creative agency costs.
And be able to say what the fee buys. If an agency cannot tell you which specific work their percentage funds, at a level more precise than “managing your account,” the number is a habit rather than a price.
When 15% is still defensible
We would rather be honest than absolute about this. There are cases:
- When the value is demonstrably outsized. If an agency can show that what they do is worth more than the fee, the percentage is beside the point. The burden of proof is on them, and it should be specific.
- When the scope genuinely is everything. Full creative production, media across several channels, measurement infrastructure, and a team that would cost more to hire.
- When spend is small. Fifteen percent of $20,000 is $3,000, which does not fund much of anything. The problem with 15% is at the top of the range, not the bottom.
The test is simple. Ask what the percentage pays for, then ask which parts of that the platform now does by itself. If most of the answer is targeting and reporting, you are paying 2015 prices for 2026 software.
This post is part of our guide to how to choose a mobile UA agency.
Frequently asked questions
What percentage do user acquisition agencies charge? The published industry range is 10-20% of media spend, with 15% common on larger accounts and the percentage usually stepping down as budgets grow. That range was set when targeting and analysis were manual work. Both have since been substantially automated, so a flat percentage at the top of that range is much harder to justify than it was. Ours is under 5%, improving with scale and with the number of channels you run.
Why is the 15% agency fee model dying? Because the work it was priced against has been automated from both directions. Algorithmic targeting absorbed manual audience building and much of the analysis, and AI production collapsed the cost of making creative. The two biggest costs inside the old fee got cheaper to deliver; the fee often did not.
Should agency fees be a percentage of spend or a flat retainer? Either can be fair. What matters is whether the fee tracks the work. A percentage that never falls as spend rises does not, because creative and strategy work does not scale linearly with budget. If you use a percentage, it should step down as you scale.
What should an agency fee actually cover in 2026? The work automation did not take: deciding what to make, curating what is worth running, testing discipline, and knowing when the platform’s delivery decisions are wrong. Those are judgment tasks, and they are what is left worth paying for.
Is 15% ever worth paying? When the scope is genuinely everything, when spend is small enough that the percentage is a modest absolute number, or when the agency can show specific outsized value. On a large account, with a narrow scope, it is very hard to defend.
